WO2023110792A1

WO2023110792A1 - Receiver, transceiver system and associated receiving method

Info

Publication number: WO2023110792A1
Application number: PCT/EP2022/085484
Authority: WO
Inventors: Raphaël LE BIDAN
Original assignee: Institut Mines Telecom
Priority date: 2021-12-17
Filing date: 2022-12-13
Publication date: 2023-06-22
Also published as: FR3131152A1

Abstract

The receiver comprises: - sampling designed to provide one or more samples (yn) per received symbol, the symbols belonging to a predefined alphabet; - an equalizer (210) designed to calculate, for each received symbol, an estimate (zn) of this symbol based on a linear combination (y'n) of the samples (yn) for this symbol; and - a decision module (216) designed to determine the symbol of the alphabet closest to the estimate (zn) as the detected symbol. The alphabet exhibits misalignment such that transmitted symbols have a non-zero predefined expected value, and the equalizer (210) is designed to add a non-zero scalar component (θ) to the linear combination (y'n) to at least partly compensate for the misalignment.

Description

TITLE: RECEIVER, TRANSMISSION/RECEPTION SYSTEM AND ASSOCIATED RECEPTION METHOD

Technical field of the invention

The present invention relates to the field of digital data transmissions and more particularly equalization for digital communications.

[0002] It relates more particularly to a receiver, a transmission/reception system comprising said receiver, an associated reception method.

Technology background

[0003] Conventionally, a digital receiver on a transmission chain comprises an equalizer to reduce the errors in detecting the symbols transmitted by a transmitter, in the presence of noise and interference between symbols (ISI: "InterSymbol Interference" in English ).

[0004] Interference between symbols may be due to:

• a limited passband of certain components of the transmission chain, e.g. photodetector of the receiver, attenuating or shifting more strongly certain frequencies of the modulated signal than others;

• a superposition at the receiver level of several echoes of the modulated signal received with proper and distinct attenuations, delays and phase shifts, typically in the case of a multipath transmission channel.

In known manner, the receiver comprises:

• a sampler designed to sample a received signal having propagated in a propagation channel, successive symbols being encoded in this received signal, in order to provide one or more samples per symbol received, the symbols belonging to a predefined alphabet;

• an equalizer designed to calculate, for each symbol received, an estimate of this symbol from a linear combination of the samples for this symbol; And a decision module designed to determine the symbol of the alphabet closest to the estimate as the detected symbol.

[0006] The article entitled "MMSE Decision-Feedback Equalizers and Coding - Part I: Equalization Results" by John M. Cioffi and M. Vedat Eyuboglu, published in IEEE Transactions on Communications, Vol. 43, No. 10, in October 1995 describes an example of a decision feedback equalizer minimizing Mean-Squared-Error. This type of filter is commonly referred to as MMSE-DFE (“Minimum Mean-Squared-Error Decision-Feedback Equalizer”).

[0007] The inventors have observed that a receiver using this type of equalizer can have reduced reception performance (e.g. high bit error rate).

[0008] It may thus be desirable to provide a receiver that improves reception performance.

Summary of the invention

[0009] A digital data receiver is therefore proposed comprising:

• an equalizer designed to calculate, for each symbol received, an estimate of this symbol from a linear combination of the samples for this symbol; And

• a decision module designed to determine the symbol of the alphabet closest to the estimate as the detected symbol; characterized in that the alphabet has an offset such that the transmitted symbols have a predefined non-zero expectation, and in that the equalizer is arranged to add a non-zero scalar component to the linear combination to at least partially compensate for the decentering.

[0010] Indeed, the non-centering of the modulation alphabet induces a bias in the estimate. In particular, with a non-centered alphabet, the equalizer MMSE-DFE of the prior art recalled above is not optimal in the sense that it does not make it possible to minimize the mean squared error criterion. [0011] The addition of the non-zero scalar component before the decision is made makes it possible to compensate at least in part for the induced bias and thus to obtain more efficient detection of the symbols, in particular by achieving a reduced bit error rate.

[0012] It follows that the decision module which follows the equalizer can be made up of decision thresholds or regions whose value or form are independent of the equalizer, and function solely of the modulation alphabet. Its practical implementation is therefore simplified, for example compared to the decision module of the receiver of the state of the art described above, where the decision thresholds must be adapted to the value of the coefficients of the equalizer.

The invention may further include one or more of the following optional features, in any technically possible combination.

[0014] Optionally, the equalizer is designed to calculate the scalar component from coefficients of the linear combination of the samples and the predefined expectation.

[0015] Also optionally, the equalizer is designed to calculate the scalar component from a channel matrix representative of the propagation channel.

[0016] Also optionally, said channel matrix is a block Toeplitz matrix.

[0017] Also optionally, the equalizer is designed to calculate said scalar component in order to minimize a mean square error between the transmitted symbols and the estimates.

[0018] Also optionally, the equalizer comprises a direct filter, preferably with a finite impulse response, and the equalizer is designed to apply this direct filter to the samples to provide said linear combination.

Also optionally, the equalizer is designed to calculate, for each symbol received, the estimate of this symbol independently of previously detected symbols.

[0020] Also optionally, the equalizer is designed to calculate, for each symbol received, the estimate of this symbol from a difference between the linear combination of the samples for this symbol and a linear combination previously detected symbols, the scalar component being added to this difference.

[0021] Also optionally, the equalizer further comprises a so-called feedback filter, preferably with a finite impulse response, and the equalizer is designed to apply this feedback filter to the previously detected symbols so as to provide said linear combination of previously detected symbols.

[0022] Also optionally, the equalizer is designed to calculate the scalar component from the direct filter.

[0023] Also optionally, the equalizer is designed to additionally calculate the scalar component from the feedback filter.

[0024] Also optionally, the equalizer is designed to iteratively update the scalar component as a function of symbols previously determined by the decision module.

[0025] A digital communication system is also proposed, characterized in that it comprises a transmitter designed to transmit symbols selected from an alphabet of non-zero expectation, and a receiver according to the invention.

[0026] A method for receiving digital data is also proposed, comprising:

• for each of several successive symbols encoded in a received signal having propagated in a propagation channel, the symbols belonging to a predefined alphabet, a reception of several samples;

• a linear equalization by calculating, for each symbol received, an estimate of this symbol from a linear combination of the samples for this symbol; And

• a determination of the symbol of the alphabet closest to the estimate as the detected symbol; characterized in that the alphabet has an offset such that the transmitted symbols have a non-zero predefined expectation, in that the linear equalization includes adding a non-zero scalar component to the linear combination to partially compensate decentering. [0027] It is also proposed a computer program downloadable from a communication network and / or recorded on a computer-readable medium, characterized in that it comprises instructions for the execution of the steps of a method according to the invention, when said program is executed on a computer.

Brief description of figures

The invention will be better understood using the following description, given solely by way of example and made with reference to the accompanying drawings in which:

• Figure 1 schematically illustrates a communication system implementing the present invention;

• Figure 2 schematically illustrates a receiver according to the invention;

• Figure 3 schematically illustrates an example of sampling performed by the receiver according to the invention;

• Figure 4 schematically illustrates an example of implementation of a receiver equalizer according to a first embodiment;

• Figure 5 schematically illustrates a process for configuring the equalizer of Figure 4;

• Figure 6 schematically illustrates an example of implementation of the equalizer of a receiver according to a second embodiment;

• Figure 7 schematically illustrates a method of configuring the equalizer according to an alternative embodiment of the first embodiment;

• Figure 8 schematically illustrates a method of configuring the equalizer according to an alternative embodiment of the second embodiment;

• Figure 9 schematically illustrates an example of implementation of the equalizer according to a variant embodiment of the second embodiment; And

• Figure 10 schematically illustrates an information processing device intended to implement the invention.

Detailed description of the invention Subsequently, the vectors will be indicated in bold type so as to distinguish them from scalars. The exponent ^H associated with a matrix will subsequently designate the conjugate transpose of this matrix.

[0030] Referring to Figure 1, an example of a digital communication system 100 in which the invention is implemented, will now be described.

The digital communication system 100 comprises a transmitter 102, a receiver 104 and a propagation channel 106 connecting the transmitter 102 to the receiver 104.

More specifically, the transmitter 102 is designed to transmit a signal e(t) carrying a digital message. By definition, a digital message is a series of binary elements, commonly referred to as "bits".

The transmitter 102 is designed to successively encode, in the signal e(t), symbols s _n respectively representing blocks or words of successive p bits. Each symbol s _n corresponds to one or more modulations of a carrier signal at a carrier frequency, such that the modulated carrier signal forms the signal e(t). Each symbol s _n belongs to an alphabet Q with K elements {S _o , Si,...,S _K -i} respectively associated with the different possible combinations of bits in a word. The transmitter 102 is designed to regularly transmit the symbols according to a symbol period Ts.

In the context of the present invention, this alphabet Q is not centered, that is to say that the transmitted symbols s _n have a non-zero expectation, ie E[s _n ] ¥= 0, in particular when the transmitted data results in a substantially equiprobable selection of the elements of the alphabet Q.

Among the modulations having a non-centered alphabet are intensity modulations, such as all-or-nothing modulation (OOK: "On-Off Keying" in English) or else its K-ary generalizations (K being the number of possible intensity levels). Any other modulation can also be used, as long as the alphabet used is not centred.

In general, it is assumed that the transmitted symbols are mutually decorrelated, that is to say that the value of a transmitted symbol does not depend on the value of the previously transmitted symbols, and that the transmitted symbols s _n have non-zero variance: o = E[|s„| ² ] - E ² [s„]. [0037] For example, in the case of an optical communication system implementing all-or-nothing (OOK) optical modulation, the transmitter 102 may be an amplitude modulated laser source configured to emit a laser pulse of amplitude A at each symbol associated with bit “1” but no pulse (ie substantially zero amplitude) at each symbol associated with bit “0”. The reverse association could also be used. In this case, the transmitted symbols s _n take their values equiprobably in the binary alphabet {0, A}, so that their expectation E[s _n ] is equal to A/2 and their variance a ² is equal to At ^2/4 .

In general, the propagation channel 106 is a material medium suitable for transmitting the signal e(t) transmitted by the transmitter 102, for example in the form of electromagnetic waves (i.e. optical or radiofrequency), so that 'it can be received by the receiver 104. For example, the propagation channel 106 is an optical fiber, a coaxial electric line, air or water or any other natural medium suitable for transmitting, for example, an optical signal or radio frequency.

The propagation channel 106 is linear in the sense that the effects or deformations that it induces on the signal are linear. For this purpose, the transmitter 102 is for example configured to transmit at a power lower than a nonlinear distortion threshold of the propagation channel 106, for example a nonlinear dispersion threshold.

[0040] For example, in the case of optical communication where the propagation channel 106 is an optical fiber, the optical power emitted is adapted according to the optical fiber used to avoid the non-linear effects of the fiber.

In addition, the propagation channel 106 is of a dispersive nature, that is to say capable of generating interference between symbols. For example, this is the case of a chromatic dispersion optical fiber inducing a broadening of the optical pulses during propagation or of a multipath radio or optical transmission channel in which several copies of the same symbol are detected. by the receiver 104 with different delays.

Furthermore, the propagation channel 106 can be marred by noise w(t), for example of zero mathematical expectation and independent of the transmitted symbols s _n . For example, the noise w(t) is a centered Gaussian additive noise with variance o _w ² . The receiver 104 is designed to receive an analog signal r(t) corresponding to the transmitted signal e(t) after propagation through the propagation channel 106.

Conventionally, the receiver 104 is for example configured to receive the received signal r(t) on the carrier frequency of the signal from the transmitter 102 and to transpose this signal into baseband (i.e. around a zero frequency) and filter it for example by applying a low-pass filter. The signal after transposition and filtering is always denoted r(t).

Knowing the alphabet Q, the receiver 104 is also designed to detect in this received signal r(t) the symbols s _n initially transmitted, in particular in the presence of noise and interference between symbols in the received signal r (t). Thus, the receiver 104 is designed to provide a detected symbol s _n for each symbol s _n transmitted.

[0046] With reference to FIG. 2, a first embodiment of the receiver 104 will now be described.

The receiver 104 comprises a sampler 208 designed to sample the received signal r(t) and thus supply successive samples y _n , _m of this signal r(t).

In particular, for each symbol period T _s corresponding to the transmission of a symbol s _n , the sampler 208 provides a set y _n of M samples y _n , _m (M greater than or equal to one) such that :

[Math 1]

where m is an integer index varying from 0 to M-1, M designating the total number of samples per symbol received during the symbol period T _s .

FIG. 3 schematically illustrates an example of received signal r(t) and the corresponding digitized signal y _n as supplied at the output of sampler 208.

In general, the sampler 208 is configured to take M samples per symbol with a sampling period T _e between each sample, such that T _e =T _s /M.

In the example of FIG. 3, the sampler 208 is configured to take two samples (M=2) per symbol period T _s . The period d sampling T _e is equal to half the symbol period T _s , so that T _e =T _s /2. Thus, the sampler 208 provides respectively for the three consecutive symbols s _n , s _n+ i, s _n+2 , the following three sample vectors:

[Math 2]

Returning to Figure 2, the receiver 104 further comprises an equalizer 210 connected to the output of the sampler 208. The equalizer 210 is configured to provide each symbol period T _s , an estimate z _n of the symbol received from the M samples y _n , _m , provided by the sampler 208.

The receiver 104 further comprises a noise estimator 212 connected on the one hand to the output of the sampler 208 and on the other hand to the equalizer 210. Conventionally, the noise estimator 212 is configured to calculate the variance o _w ² of the noise from the samples y _n supplied by the sampler 208 and supply this variance to the equalizer 210.

In the present example, the noise is an additive Gaussian noise, centered (ie of zero expectation), of variance o _w ² and independent of the transmitted symbols. More generally, the noise characteristics are specific to the communication system concerned and depend in particular on the nature of the transmission channel 106. Thus, these characteristics can be adjusted according to the transmission system considered.

The receiver 104 further comprises a channel estimator 214 connected on the one hand to the output of the sampler 208 and on the other hand to the equalizer 210. The channel estimator 214 is configured on the one hand to calculate from the samples y _n , a channel matrix H representative of the deformations that the propagation channel 106 causes the symbols to undergo and on the other hand to supply this channel matrix H to the equalizer 210. This matrix describes the model interference between symbols specific to the propagation channel concerned.

Preferably, the channel matrix H is a Toeplitz matrix by blocks, comprising for example at least one non-zero block in the first position of the first row and first column. The equalizer 210 is configured to receive data relating to the modulation format used by the transmitter 102 to transmit the symbols. For example, these data include the expectation E[s _n ] and/or the variance a ² of the symbols s _n transmitted.

The equalizer 210 will be described in more detail later, with reference to Figure 4.

The receiver 100 further comprises a decision module 216 designed to select the symbol of the alphabet Q closest to the estimate z _n . The selected symbol is thus taken as the detected symbol s _n- , where A is the restitution delay which is a parameter of the equalizer 210.

For example, the decision module 216 is designed to compare the estimate z _n with regions defined by predefined thresholds, each region being respectively associated with one of the symbols of the alphabet Q. The thresholds are independent of the equalizer 210 and in particular of its coefficients p ^H and of its scalar component 0 which will be described later.

[0061] Referring to Figure 4, the equalizer 210 of the receiver 104 of Figure 2 will now be described in more detail.

The equalizer 210 comprises a sample combiner, called direct combiner 400, designed, for each symbol period T _s , to receive the samples y _n and provide a linear combination y′ _n of said samples.

For example, the direct combiner 400 is defined by direct coefficients p ^H , so that the linear combination y' _n is obtained by weighting the samples y _n by the direct coefficients p ^H , so that y' _n = p ^H xy _n , where x is the dot product.

For example, this direct combiner 400 is a direct filter ("feed-forward" in English), preferably a digital filter with finite impulse response (FIR: "Finite Impulse Response" in English).

The equalizer 210 further comprises a configuration module 400a of the direct combiner 400, to configure the latter according to configuration parameters.

For example, these configuration parameters include channel parameters specific to the propagation channel 106, such as the channel matrix H and/or the noise variance o _w ² . These configuration settings may also include parameters specific to the modulation alphabet Q, such as the variance o _s ² of the symbols s _n transmitted.

In the case where the direct combiner 400 is an Fl R filter, the configuration module 400a determines the coefficients p ^H defining the filter. For example, this vector p ^H comprises N×M coefficients, so that the direct combiner 400 provides a linear combination of samples every T _s seconds.

The equalizer 210 further comprises a feedback loop between the decision module 216 and the direct combiner 400, in order to improve symbol detection from previously detected symbols. This loop comprises on the one hand a combiner of detected symbols, called reverse combiner 420, at the output of the decision module 216 and on the other hand a subtractor 440 between the direct combiner 400 and the reverse combiner 420.

The reverse combiner 420 is configured to supply, for each symbol period T _s , a linear combination of the symbols s _nA-1 , s _nA-2 > ■■■ previously detected by the decision module 216 A symbol periods earlier . This linear combination of symbols is denoted s' _n .

For example, the reverse combiner 420 is a backward filter, preferably a digital filter with finite impulse response (F1). Alternatively, the feedback filter could be an infinite impulse response (HR) filter.

The equalizer 210 further comprises a configuration module 420a of the reverse combiner 420 designed to configure the latter according to parameters specific to the propagation channel 106, such as the channel matrix H, and configuration parameters of the combiner direct 400, such as the p ^H coefficients.

In the case where the feedback filter is an FIR filter, the configuration module 420a of the feedback combiner 420 of symbols determines a vector of coefficients q ^H defining the filter, this vector comprising N _b coefficients, so that the filter feedback 420 provides a linear combination of samples every T _s seconds.

The subtractor 440 is configured to calculate a difference d between the linear combination y′ _n of the samples y _n and the linear combination of the previously detected symbols s′ _n , ie the difference d=y − s′ _n .

The equalizer 210 further comprises an addition module 460 between the subtractor 440 and the decision module 216. Advantageously, the addition module 460 is configured to add a nonzero scalar component 0 to the difference d, so that the estimate z _n is given by z„ = y,' - s'„ + 0.

The addition of this non-zero scalar component 0 makes it possible to compensate at least in part for the bias resulting from the impact of the expectation E[s _n ] of the transmitted symbols s _n through the combined action of the matrix of channel H and coefficients p ^H .

The equalizer 210 further comprises a configuration module 460a designed to calculate the non-zero scalar component 0 and to supply this component to the addition module 460 so as to add it to the difference d.

For example, the configuration module 460a is configured to calculate the non-zero scalar component 0 as a function of parameters linked to the alphabet Q, such as the expectation E[s _n ] of the transmitted symbols s _n , of parameters configuration of the upstream combiner 300, such as the coefficients p ^H , and/or parameters specific to the propagation channel 106, such as the channel matrix H.

Thus, in the case where the equalizer 210 implements an upstream filter 300 FIR defined by the coefficients p ^H and a return filter 320 FIR defined by the coefficients q ^H , the estimate z _n provided at the input of the module of decision 216 is such that:

[Math 4] z„ = p ^{H x} y _n - q ^H xs _n + 0

( ^s n-4-1 \ j denotes the vector grouping the N _b past decisions relating nA-Nb 'to the symbols transmitted A+1 symbol periods earlier.

The configuration modules 400a, 420a, 460a are designed to respectively configure the direct combiner 400, the return combiner 420 and the addition module 460 of the scalar component. These modules can be implemented in software form executed by the same information processing module, such as that described below with reference to Figure 9.

[0080] With reference to FIG. 5, a method 500 of configuring the equalizer 210 will now be described, in the case where the upstream combiner 400 and the reverse combiner 420 are FIR filters as described with reference to FIG. 4.

During an initialization step E50, the receiver 104 receives from the transmitter 102 a known reference message. The noise estimator 212 and the channel estimator 214 previously described with reference to FIG. 2 determine from a analysis of this known message, respectively the variance of the noise o _w ² and the channel matrix H and supply this information to the equalizer 210.

During a sampling step E51, the sampler 208 samples the received signal r(t) corresponding to an unknown message e(t) transmitted by the transmitter 102, so as to provide a set of blocks of M successive samples for N received symbols. Thus, a block of M samples is supplied every T _s seconds (symbol period) by the sampler 208.

By collecting the last NxM samples received (i.e. N blocks of M samples) in the form of vectors, up to the time NT _S +(M-1) T _s /M, the samples can be written in the modeling framework presented as follows: [Math 5]

H denoting the channel matrix, preferably equal to a block Toeplitz matrix describing the inter-symbol interference model specific to the propagation channel, s _n being the vector of transmitted symbols participating in the inter-symbol interference present in the NxM last samples received, and w _n is a Gaussian noise vector, centered, of variance o _w ² on each coordinate and independent of the transmitted signal.

For example, the channel matrix comprises two bare blocks, more particularly a lower null triangular block and an upper null triangular block.

During a calculation step E53, the configuration module 400a of the forward combiner 400 calculates the coefficients p ^H , the configuration module 420a of the reverse combiner 420 calculates the coefficients q ^H and the configuration module 460a of the module d addition 460 calculates the nonzero 0 scalar component.

For example, the parameters p ^H , q ^H , 0 of the equalizer 210 are calculated so as to minimize the “mean squared error” cost function (MSE: “Minimum Square Error”) between the estimated z _n and the symbol s _n.Ai where A designates the delay in restitution of the symbols by the receiver, denoted:

[Math 6]

Alternatively, the parameters p , q ^H , 0 of the equalizer 210 are calculated so as to minimize a bit error rate (BER: “Bit Error Rate”).

Advantageously, the delay A can be optimized. For example, it is possible to try several values of the delay A (for example by scanning) to find the best one, i.e. the one minimizing the cost function J.

The result of this function J minimization calculation makes it possible to determine the parameters p ^H , q ^H , 0 of the equalizer 210 according to the following expressions:

[Math. 7]

values of the columns of the channel matrix H! ^A+I = H(: , A + 1) denotes column A+1 of channel matrix H; E(s _n ) denotes the expectation of the transmitted symbols; o _s ² denotes the variance of the transmitted symbols; q* denotes the coefficient of rank i in q ^H ; I denotes the identity matrix; J _A = with s=N+L-2-AN _b ; and 0 denotes the zero matrix.

Thus, the value of non-zero scalar component 0 is defined in particular as a function of the mathematical expectation of the non-centered alphabet Q and of the channel matrix H modeling in particular the interferences between symbols.

This result is obtained by making the following assumptions: the value of the symbols s _n transmitted is a stationary random variable, so that the mathematical expectation of the symbols is invariant over time, ie E(s _n ) = E( s _nA ); the noise is decorrelated from the symbols s _n emitted, so that the covariance of the noise and of the symbols is zero; and the noise has zero mathematical expectation, ie E(w _n ) = 0.

The expressions calculated above are obtained assuming Gaussian white noise with zero mean. However, these could of course be adapted by those skilled in the art, so as to take account of other noise statistics, for example correlated noise, while remaining within the scope of the present invention. The equations [Math 7] are for example determined beforehand and implemented respectively in the configuration modules 400a, 420a, 460a.

During a configuration step E55, the parameters p ^H , q ^H , 0 are applied respectively to the upstream filter 300, to the return filter 320 and to the addition module 360 so as to configure the equalizer 210.

[0094] As soon as the equalizer 210 is configured, the configuration parameters remain unchanged for the reception of several symbols.

With reference to FIG. 6, a second embodiment of the receiver according to the invention will now be described.

According to the second embodiment, the equalizer, which now bears the reference 610, differs from the equalizer 210 according to the first embodiment, in that it does not include a feedback loop.

Thus, the direct combiner 400 supplies the linear combination y _n from the samples y _n and the addition module 460 adds to this linear combination y _n the non-zero scalar component 0 so as to obtain the estimate z _n , such that z _n =y _n +0. In this way, the estimate z _n is calculated independently of the symbols Sn-â-1, Sn-â-2, ■■■ previously detected.

As described above, the direct combiner 400 can be an FIR filter configured by the configuration module 400a is designed to calculate the coefficients p ^H . Thus, the estimate z _n is determined according to these coefficients, such that z _n = p ^H xy _n + 0.

The method described with reference to FIG. 5 remains valid except that only the coefficients p ^H and the non-zero scalar component 0 are calculated during step E53.

The parameters p ^H and 0 of the equalizer 610 are calculated by making the same assumptions as those previously for the first embodiment. In this case, the parameters of the equalizer 610 are determined according to the following expressions, with the same notations as previously:

[Math. 8]

It can be seen that the non-zero scalar component 0 depends solely on the predefined expectation of the alphabet Q, of the channel matrix H and of the coefficients P ^H -

[0102] Assuming that the value of the symbols s _n is a stationary random variable then the associated expectation is invariant over time, so that E(s _nA )=E(s _n ). ^Assuming that the propagation channel 106 is stationary over a reference period, the channel matrix H is time invariant. In this case, the non-zero scalar component O is a constant over this reference period.

In the two embodiments described above, the parameters of the equalizer 210 are initially calculated analytically, in particular as a function of external parameters linked to the propagation channel 106, to the noise and to the statistical properties of the symbols s _n . In particular, the channel estimator 214 and the noise estimator 212 can be used to supply the equalizer 210 with the channel matrix H and the noise variance o _w ² .

[0104] Alternatively, the parameters (0, p ^H , q ^H ) or (0, p ^H ) of the equalizer 210 can be determined adaptively by a technique of convergence and tracking. This technique constitutes a variant embodiment that can be applied to each of the embodiments described above.

[0105] With reference to FIG. 7, a variant of the first embodiment of the equalizer will now be described for determining and adjusting the parameters of the equalizer according to a convergence and tracking technique.

According to this variant, the equalizer, which now bears the reference 710, is designed to iteratively update the scalar component 0 as a function of symbols previously determined by the decision module 216. This does not require use a decoder (“soft demapper” in English) which uses several successive estimates, to analyze them as with error correction. On the contrary, the decision module 216 provides each symbol s _n- from a single estimate z _n .

Thus, the equalizer 710 further comprises a module 780 for comparing the symbols previously detected by the decision module 216 with the estimate z _n , so as to calculate an error signal e _n . For example, the error signal e _n provided at the output of the comparison module 780 is such that e _n =z _n −s _nA . The comparison module 780 has an output connected to the configuration module 700a of the direct combiner 400, to the configuration module 720a of the reverse combiner 420 and to the configuration module 760a of the add module 460, so that the signal of The error e _n is supplied, for each symbol period, simultaneously at the input of these configuration modules 700a, 720a, 760a.

[0109] A method 800 of configuring the equalizer 710 will now be described in more detail with reference to Figure 7.

During an initialization step E80, the parameters p ^H , q ^H , 0 of the equalizer 710 are initialized, for example with the following default values: p ^H =(1,0,0,. ..,0); q ^H = (0,0,0,...,0); 0 =0.

During a transmission/reception step E82, the transmitter 102 starts sending a sequence of known symbols and the receiver 104 receives this sequence.

This known sequence of symbols is used by the equalizer 710 to update its configuration parameters p ^H , q ^H , 0, as it receives the symbols of the known sequence. This update is carried out iteratively, during a convergence phase B1 as detailed below.

At each new set of M samples y _n supplied by the sampler 208, the configuration parameters p ^H , q ^H , 0 of the equalizer 710 are updated. Thus, the parameters of the equalizer 710 are updated once per symbol period, for each symbol of the known sequence, received during the convergence phase B1.

This updating is carried out by comparing the estimates z _n of the symbols with the symbols of the known sequence.

Each symbol s _n of the known sequence is received by the receiver with a delay A. Thus, during a comparison step E84, the comparison module 780 supplies, for each symbol period, the error signal e (n) between the estimate z _n and the received symbol s _nA such that e _n = z _n -s _n-& . This error signal e(n) is transmitted at each symbol period to the configuration modules 700a, 720a, 760a.

During an update step E86, the configuration modules 700a, 720a, 760a calculate the configuration parameters p ^H , q ^H , 0 of the equalizer 710 according to the initial value fixed during the initialization step E80 or a previous value of the parameters determined during the previous period (ie during a previous iteration). [0117] For example, during the update step E86,

• the configuration module 700a of the upstream filter 400 is configured to calculate the coefficients p ^H of said filter according to the following recurrence formula:

[Math 9]

Pn+l = Pn ~ Pie _n yn

• the configuration module 720a of the feedback filter 420 is configured to calculate the coefficients q ^H of said filter according to the following recurrence formula:

[Math 10]

• the configuration module 760a of the addition module 460 is configured to calculate the scalar component 0 according to the following recurrence formula:

[Math 11 ]

©n+l = ®n ~ 2e _n where e _n = z _n - s„_ _A designates the difference (or error) between the estimate z _n and the symbol s _n . A transmitted with a delay A for the symbol period T _n ; p! denotes an adaptation step, preferably less than 1, for the adaptation of the coefficients p ^H of the upstream filter and the coefficients q ^H of the return filter and p ₂ denotes an adaptation step, preferably less than 1, for the 'adaptation of the scalar component 0. For example, the adaptation steps p! and p ₂ are equal, so that pi=p ₂ .

This updating is carried out iteratively, ie period by period, so as to converge towards a target value. Thus, after each update, the equalizer 710 determines during a test step E88, whether a convergence criterion has been reached. For example, a convergence criterion is such that the difference e _n is less than a predetermined convergence threshold e^

As soon as the error e _n is less than the predetermined convergence threshold ei, (ie, e(n)<ei) a tracking phase B2 is implemented which will be described below in more detail. During this tracking phase B2, the updating of the settings of the equalizer 710 is controlled by the previous detections of symbols. Thus, the tracking step B2 differs from the convergence step B1 in that it uses the detected symbols s„_ _A in the error e _n instead of the transmitted symbols s _nA to update the parameters of the 710 equalizer.

[0121] Thus, during a step E89, the error signal e(n) is replaced by an error signal between the estimate z _n and the symbol s„_ _A detected with a delay A so that

The parameters of the equalizer 710 are determined at each iteration during the update step E86, by the same equations Math 9, Math 10 and Math 11 as described above, as long as the error _in = z„ - s„_ _A is greater than the predetermined convergence threshold e^

As soon as this error e(n) becomes greater than the predetermined convergence threshold e _{1 ;} the method switches to convergence phase B1 by executing steps E84, E86 as long as e _n >ei as described previously.

[0124] With reference to FIG. 9, a variant of the second embodiment of the equalizer will now be described for determining and adjusting the parameters of the equalizer according to the convergence and tracking technique previously described.

According to this variant, the equalizer, which now bears the reference 910, is designed to iteratively determine its configuration parameters 0, p ^H .

As described with reference to FIG. 7, the equalizer 910 comprises the module 780 for comparing the symbols previously detected by the decision module 216 with the estimate z _n , so as to calculate the error signal e _n . As previously described, the configuration parameters 0, p ^H of the equalizer 910 are updated at each iteration as a function of the error signal e(n), respectively by the configuration modules 900a and 960a according to the same method 800 as described with reference to FIG. 8 with the difference that only the parameters 0, p ^H .

For example, during the update step E86, the configuration module 900a of the upstream filter 400 is configured to calculate the coefficients p ^H of said filter according to the recurrence formula [Math 9]; the configuration module 960a of the addition module 460 is configured to calculate the scalar component 0 according to the recurrence formula [Math 11],

In the examples described above, the receiver 106 according to the invention may comprise a computer system 10, as illustrated in the form of a block diagram in FIG. 10, adapted for the implementation of one or more modules described above.

This computer system 10 comprises a data processing unit 10.1 (such as a microprocessor denoted CPU, from the English "Central Processing Unit"), a main memory 10.2 (such as a random access memory, denoted RAM, from the English "Random Access Memory") accessible by the processing unit 10.1, a read only memory 10.3 denoted ROM, from the English "Read Only Memory"), accessible by the processing unit 10.1, a support optional storage, readable by computer, such as for example a local medium (such as a local hard disk 10.6, denoted HD in English "Hard Drive") or even a removable medium (such as a USB key, the English “Universal Serial Bus”, or else a CD, from the English “Compact Disc” or else a DVD, from the English “Digital Versatile Disc”) readable by means of an appropriate reader of the computer system 10 ( such as a USB port or a CD and/or DVD disk drive); a 10.7 input/output module for receiving/sending data from/to external devices such as hard disk, removable storage medium or others.

A computer program P containing instructions in the form of an executable code for the processing unit 10.1 is recorded on the medium 10.6.

This computer program P is for example intended to be loaded into the main memory, so that the processing unit 10.1 executes its instructions. These instructions implement one or more of the modules described previously, which are thus software modules.

Alternatively, all or part of these modules could be implemented in the form of hardware modules, that is to say in the form of an electronic circuit, for example micro-wired, not involving a computer program. .

It clearly appears that a receiver such as the one described previously makes it possible, thanks to the addition of a non-zero scalar component in the estimate z _n before detection of the symbol, to improve the symbol detection performance. [0134] It will also be noted that the invention is not limited to the embodiments described above. It will indeed appear to those skilled in the art that various modifications can be made to the embodiments described above, in the light of the teaching which has just been disclosed to them.

The above embodiments and variants have been described to improve the detection performance of a digital data receiver in the presence of interference between symbols, which is the most frequent case of use. More generally, the invention can also be applied to any other type of distortion or interference caused by the propagation channel, insofar as these can be modeled in the form of a Toeplitz matrix compatible so that the sampled signal y _n can be written in the matrix form y _n =Hs _n +w _n according to the expression [Math 5].

In the examples described above, finite impulse response filters have been described to implement the sample combiner (i.e. direct filter) and the previously detected symbol combiner (i.e. return filter). Other types of filters may also be used.

In the detailed presentation of the invention which is made above, the terms used must not be interpreted as limiting the invention to the embodiments set out in the present description, but must be interpreted to include therein all the equivalents of which the forecast is within the reach of those skilled in the art by applying their general knowledge to the implementation of the teaching which has just been disclosed to them.

Claims

[1] Digital data receiver (104) comprising: a sampler (208) designed to sample a received signal (r(t)) having propagated in a propagation channel (106), successive symbols (s _n ) being encoded in this received signal (r(t)), in order to provide one or more samples (y _n ) per symbol (s _n ) received, the symbols (s _n ) belonging to a predefined alphabet (Q); an equalizer (210; 610; 710; 910) designed to calculate, for each symbol (s _n ) received, an estimate (z _n ) of this symbol (s _n ) from a linear combination (y' _n ) of the samples (y _n ) for this symbol (s _n ); and a decision module (216) adapted to determine the symbol of the alphabet (Q) closest to the estimate (z _n ) as the detected symbol (s„); characterized in that the alphabet (Q) has an off-centering such that the transmitted symbols (s _n ) have a non-zero predefined expectation (E[s _n ]), and in that the equalizer (210; 610; 710 ;910) is designed to add a non-zero scalar component (0) to the linear combination ( _y'n ) to at least partially compensate for the off-centering; the equalizer (710; 910) being designed to update (E86) iteratively (B1, B2) the scalar component (0) as a function of symbols previously determined by the decision module (216).

[2] Receiver according to claim 1, in which the equalizer (210; 610) is designed to calculate the scalar component (0) from coefficients (p ^H ) of the linear combination (y' _n ) of the samples (y _n ) and the predefined expectation ( [s„]).

[3] Receiver according to claim 2, in which the equalizer (210; 610) is designed to calculate the scalar component (0) from a channel matrix (H) representative of the propagation channel (106).

[4] Receiver according to claim 3, wherein said channel matrix (H) is a block Toeplitz matrix.

[5] Receiver according to any one of claims 1 to 4, in which the equalizer (210; 610) is arranged to calculate said scalar component (0) in order to minimize a mean square error between the transmitted symbols (s _n ) and the estimates (z _n ).

[6] Receiver according to any one of Claims 1 to 5, in which the equalizer (210; 610; 710; 910) comprises a direct filter (400), preferably with a finite impulse response, and the equalizer is designed to apply this direct filter (400) to the samples (y _n ) to provide said linear combination (y' _n ).

[7] Receiver according to any one of Claims 1 to 6, in which the equalizer (610) is designed to calculate, for each symbol (s _n ) received, the estimate (z _n ) of this symbol (s _n ) independently of previously detected symbols.

[8] Receiver according to any one of claims 1 to 6, in which the equalizer (210; 710) is designed to calculate, for each symbol (s _n ) received, the estimate (z _n ) of this symbol ( s _n ) from a difference (d) between the linear combination (y' _n ) of the samples (y _n ) for this symbol (s _n ) and a linear combination ( ) of previously detected symbols (s _n-1 , ..., s _n-2! ), the scalar component (0) being added to this difference (d).

[9] Receiver according to claim 8, in which the equalizer (210; 710) further comprises a so-called feedback filter (420), preferably with a finite impulse response, and the equalizer (210; 710) is designed to applying this feedback filter (420) to the previously detected symbols (s _n-1 , ..., s _n-2! ) so as to provide said linear combination (s') of previously detected symbols (s _n-1; .. ., _sn-2! ).

[10] Receiver according to claim 6, in which the equalizer (210; 610; 710; 910) is designed to calculate the scalar component (0) from the direct filter (400).

[1 1] Receiver according to claims 9 and 10, wherein the equalizer (210; 710) is arranged to calculate the scalar component (0) additionally from the feedback filter (420).

[12] Receiver according to any one of claims 1 to 1 1, wherein the equalizer (710; 910) is configured to calculate the scalar component (0) according to the following recurrence formula: [Math 12]

where e _n = z _n - s _n-& denotes the difference (or error) between the estimate z _n and the symbol s _n . _A transmitted with a delay A for a symbol period T _n ; p ₂ denotes an adaptation pitch, preferably less than 1, for the adaptation of the scalar component e.

[13] Digital communication system (100) characterized in that it comprises a transmitter (102) designed to transmit symbols selected from an alphabet (Q) of non-zero expectation, and a receiver (104) according to one any of claims 1 to 12.

[14] Method for receiving digital data, comprising: for each of several successive symbols (s _n ) encoded in a received signal having propagated in a propagation channel (106), the symbols (s _n ) belonging to an alphabet (Q) predefined, a reception of several samples (y _n ); a linear equalization by calculating, for each symbol (s _n ) received, an estimate (z _n ) of this symbol (s _n ) from a linear combination (y' _n ) of the samples (y _n ) for this symbol ( _sn ); and a determination of the symbol of the alphabet (Q) closest to the estimate (z _n ) as the detected symbol (s _n ); characterized in that the alphabet (Q) has an off-centering such that the transmitted symbols (s _n ) have a non-zero predefined expectation (E[s„]), in that the linear equalization comprises an addition of a non-zero scalar component (0) to the linear combination (y' _n ) to at least partly compensate for the off-centering, the scalar component (0) being updated iteratively (B1, B2) as a function of symbols previously determined by the decision module (216).

[15] Computer program (P) downloadable from a communication network and/or recorded on a computer-readable medium, characterized in that it comprises instructions for the execution of the steps of a method according to claim 14 , when said program (P) is executed on a computer.